Network Interaction
The K-Link Device does not register trends or COV requests on the target systems. Instead, it uses its own polling engine to directly query the data values following a drip series-parallel methodology. This means the engine is not pulling more than one data point at a time off each trunk of the network. As such the bandwidth or CPU utilization at any point in time is very low for the targets. The polling engine collects response times and maps out network health as it interacts with the target system(s) which has allowed CopperTree to assist in network optimization as part of its deployment.
The specific way in which the polling engine pulls data varies by target system as systems lack a long-term standard. But the general concept is that the K-Link leverages the access it is provided (some it is open by simply being on the network, and others it has been given login credentials) to query data from the target system via the relevant data exchange protocols (XML, Telnet, BACnet, HTTP, web services, etc). The biggest exception to this is the legacy Andover driver which requires its own site-specific CT optimized function to be loaded into some of the controllers.
For very slow networks (single twisted pair) with lots of equipment, the way CopperTree minimizes network footprint is by deploying multiple K-Link. By having a sufficient number of geospatially distributed Devices, the Device can avoid going across the broader network thus preserving bandwidth on that bus.
The Device checks each point based on the intersection of target polling intervals based on the function of the data stream and the network health. The Device tracks its own COV as part of the outbound data compression. This means that each Device is typically transmitting less bandwidth than a YouTube stream and the average data transfer is less than 20GB / month.
BAS Data Acquisition
The K-Link will scan and collect the current values of your BAS points every 5 minutes by default, though this can be adjusted. The polling interval may appear longer depending on the Device’s load, network speed, and various other factors that may cause the reads to take longer than 5 minutes.
All samples will be collected, except for the samples which have not changed by at least 0.5% of their previous value since the last sample. This will continue for 3 hours, then, if the value still has not changed by 0.5% we will collect a new sample.
If the absolute value of the sample is less than 5 it will be collected every 5 minutes regardless.
Samples are collected to an accuracy of 3 decimal places.
